ISO 639 - New item approved - N'Ko
mrc at CAC.Washington.EDU
Thu Jun 8 20:56:14 CEST 2006
This may sound awfully retro, but a programmer's view bears stating in
this forum of linguists.
To be truly useful, any multi-national/multi-lingual registry of names
will have the following fields:
(1) There should be a field which has the name represented in vernacular
English using exclusively characters from the ASCII character set.
Simplicity for ASCII searches is more important here.
(2) There should be a field which has the name represented in the
associated language (e.g., "Deutsch", "Deutschland" as opposed to
"German", "Germany") and which can use any/all Unicode characters.
Linguistic accuracy is more important here.
(3) There probably should be a field, associated with (2), which contains
that string represented in IPA.
(4) If it is necessary to assauge French sensibilities, it is alright to
have a field that is equivalent to (1), but permits codepoints found in
ISO 8859-1 (that is, Unicode codepoints up to U+00ff).
(5) Similar concessions to (4), but with the other "official UN
languages" (I think there are a half dozen of these).
(6) Similar concessions to (5), but only if a sufficient use case is
presented (e.g., sufficient number of native speakers). Japanese, Korean,
Hindi, etc. may make the cut. Lower Slobbovian would not, except for the
(2) entry for Lower Slobbovia and Lower Slobbovian.
(7) For all of the above, there are zero or more "nicknames" which are
With the N'Ko example, I can easily envision the (1) and (2) fields being
different, even if typographically they look more or less identical.
Personally, I feel that (4) should not be done except as part of (5) (what
makes the French more important than the Russians or the Chinese?) and
perhaps (4) should not be presented as a separate item in an open forum.
-- Mark --
Democracy is two wolves and a sheep deciding what to eat for lunch.
Liberty is a well-armed sheep contesting the vote.
More information about the Ietf-languages